Theoretical Study of the Gas-Phase Reaction of ... - ACS Publications

CEP 05513-970. Phone/Fax: 55-11-3818-3888. E-mail: jmrnigra@ iq.usp.br. † E-mail: ... trapping plates with the radio frequency field tuned to ∼7.5...
0 downloads 0 Views 96KB Size
Download PDF
J. Phys. Chem. A 2002, 106, 371-378

371

Theoretical Study of the Gas-Phase Reaction of Fluoride and Chloride Ions with Methyl Formate Josefredo R. Pliego, Jr.† and Jose´ M. Riveros* Instituto de Quı´mica, UniVersidade de Sa˜ o Paulo, Caixa Postal 26077, CEP 05513-970 Sa˜ o Paulo, SP, Brazil ReceiVed: April 16, 2001; In Final Form: October 17, 2001

The potential energy surface of the gas-phase reaction between halide ions (F- and Cl-) and methyl formate has been investigated by ab initio calculations. For F-, two pathways have been observed at thermal energies and identified in the calculations: (1) R-elimination of CO to yield a fluoride-methanol adduct, the socalled Riveros reaction that has found wide application in gas-phase ion chemistry, and (2) SN2 displacement of HCOO-. The first reaction is shown to proceed by the initial formation of a loose complex followed by formal abstraction of a formyl hydrogen to yield a three-body complex that dissociates into the final products. The SN2 reaction initially involves formation of a loose complex with the fluoride attached to the methyl group of the ester. The first pathway is calculated to go through a lower energy local transition state than the corresponding SN2 reaction but the transition states are located below the energy of the reagents. Both ionneutral complexes can interconvert via formation of a stable tetrahedral intermediate. The product distribution was estimated via a simple RRKM calculation that predicts 92% of R-elimination and 8% of SN2 reaction. This prediction is in excellent agreement with measurements carried out by FT-ICR. This product distribution is predicted to remain essentially unchanged for the reaction with DCOOCH3 in agreement with experimental observations. A similar analysis of the corresponding Cl- + HCOOCH3 reaction reveals that R-elimination has a substantial activation energy (well above the reagents) accounting for the failure to observe this reaction even though it is exothermic. These calculations also reveal that for the Cl- system, the tetrahedral intermediate is not a stable intermediate in agreement with previous experimental data on related systems.

Introduction

SCHEME 1

Gas-phase ion chemistry, and particularly the study of ionmolecule reactions, have significantly furthered our understanding of intrinsic reactivity and helped to unravel the mechanism of chemical reactions under solvent-free conditions.1 This is best exemplified by gas-phase nucleophilic displacement (SN2) reactions where a lively interplay between experiment and theory has contributed to the elucidation of the energetics, the potential energy surface, and the dynamics of these reactions.2-4 Because of their relevance to many important organic and biochemical processes gas-phase reactions between esters and anionic nucleophiles have also been the subject of a large number of experimental studies.5 However, considerably less is known about the potential energy surface and the dynamics for these systems since several mechanisms leading to different products are operative even for the simplest cases. Experiments from our laboratories6,7 as well as work from other groups5d,5f,8,9 have revealed that the reaction of ions such as OH-, F-, RO- (R ) alkyl), and NH2-, with methyl formate can proceed by R-elimination of CO, SN2 displacement at the methyl group, and/or addition-elimination at the carbonyl center as illustrated in Scheme 1 (X- ) nucleophile). While the product distribution is dependent on the thermochemistry and the nature of the nucleophile, reaction 1a is the preferred pathway for ions such as F-, OH-, and RO-. As a * Author to whom correspondence should be addressed at Instituto de Quı´mica, University of Sa˜o Paulo, Caixa Postal 26077, Sa˜o Paulo, Brazil, CEP 05513-970. Phone/Fax: 55-11-3818-3888. E-mail: jmrnigra@ iq.usp.br. † E-mail: [email protected].

result, reaction 1a has become well established in gas-phase ion chemistry10 and has been routinely used as a convenient method to generate gas-phase mono-solvated anions11 in experiments designed to determine binding energies,12 structure,13 and spectroscopic characterization.14 On the other hand, it is still unclear as to why reaction 1a is not observed for second-row anions such as Cl- although the reaction is exothermic. Given the ease of reaction 1a for simple nucleophiles, the question arises regarding the nature of the potential energy surface that favors such a pathway as opposed to the traditional viewpoint of nucleophilic attack at the carbonyl system. We have recently addressed this particular question for the reaction of OH- vis-a`-vis the gas-phase hydrolysis reaction 1b that can proceed either by a BAC2 or an SN2 mechanism.15 At the highest level of theory reported in our earlier work (MP4/6-311+G(2df,2p)//MP2/6-31+G(d)), reaction 1a was found to proceed by abstraction of the formyl proton and formation of a threebody complex, CH3O-(H2O)(CO) without any formal activation energy along the reaction pathway. In this report, we extend our ab initio calculations to the F- and Cl- reactions in order to understand two important features: (a) the strong preference for reaction 1a over the competing SN2 reaction 1b for F-; (b) the lack of any reactivity for Cl-.

10.1021/jp0114081 CCC: $22.00 © 2002 American Chemical Society Published on Web 12/20/2001

372 J. Phys. Chem. A, Vol. 106, No. 2, 2002

Pliego and Riveros

Failure to observe a reaction between Cl- and HCOOCH3 was originally rationalized in terms of the thermochemistry involved in the association of anions with methanol and the energy required to decarbonylate methyl formate.16 Unfortunately, this argument cannot explain the origin of a possible energy barrier for the case of Cl-. Our present calculations for the energy surface profile of the F- reaction coupled with statistical rate theories yield excellent agreement with the experimentally observed product distribution. Likewise, these calculations reveal a substantial energy barrier for reaction 1a when Cl- is the nucleophile. Thus, it is possible to have a unified view about this useful and common reaction in the gas phase. Experimental Section A recent study5f of the gas-phase F- + HCOOCH3 reaction showing a considerably different product distribution from our earlier experiments7 prompted us to reexamine this system in our FT-ICR spectrometer. The general features of this spectrometer have been discussed in recent publications from this laboratory.17 The temperature of the cell is typically around 330 K while the magnetic field was maintained at 1.0 T. Fluoride ions were generated directly in the cell of the spectrometer by electron impact at 3.4 eV from NF3 at a nominal pressure of 1.8 × 10-8 Torr. For a typical 80 ms ionization pulse, an electron ejection pulse of 150 ms was applied to the trapping plates with the radio frequency field tuned to ∼7.5 MHz. The electron ejection pulse is essential for reducing space charge effects due to trapped thermal electrons. Methyl formate (Aldrich) was introduced in the spectrometer at a pressure of 7.5 × 10-9 Torr, and its pressure was periodically remeasured after pumping away NF3. Nitrogen was added to the system up to a total pressure of 2 × 10-7 Torr in order to help the thermalization of the reagent ions. In a different set of experiments, nitrogen was introduced through a pulsed valve to achieve a maximum instantaneous pressure of 5 × 10-7 Torr. Ion/molecule reactions were studied by isolating fluoride ions in the reaction cell through a sequence of short pulses to remove all other ions at times ranging from 500 to 2500 ms after ion formation. The reaction products of the F-/HCOOCH3 system were determined from FT-ICR spectra typically recorded with 32 K data at different reaction times. While no attempt was made to measure the absolute rate constants in these experiments due to the uncertainties in the value of the absolute pressures, the product distribution was determined from the relative peak intensities of the magnitude spectra of the product ions18 acquired with impulse excitation.19 Impulse excitation is routinely used in our laboratory because at low ion densities it yields excellent isotopic ratios as frequently verified in our spectrometer with germanium-containing ions.17d The advantage of impulse excitation in these cases resides in the fact that all ions in the cell are accelerated simultaneously to the same large radius of gyration in a short period of time thus minimizing some of the problems associated with mass discrimination in FT-ICR spectra.20,21 F- reacts with HCOOCH3 along two different pathways as shown in reaction 2 along with the product distribution. These

results represent the average of 10 independent measurements and were found to be independent of reaction time within

Figure 1. FT-ICR spectra of the products of reaction 2. The left-hand side shows the product distribution for the F-/HCOOCH3 reaction and the right side the corresponding product distribution for the F-/DCOOCH3 reaction.

experimental error. Similar experiments were also carried out with 99% DCOOCH3 and the product distribution is observed to be essentially unchanged upon deuteriation at the formyl hydrogen (see Figure 1). These results are in excellent agreement with those reported in our previous work.7Experiments carried out with a 10-fold, or higher, increase in the pressure of methyl formate reveal that F-(HOCH3) ions can undergo a slow reaction similar to reaction 2a to yield F-(CH3OH)2 ions. This kind of sequential “Riveros reaction” initiated by a mono-solvated anion has already been observed for alkoxide ions,22 and for CF3-.12c Ab Initio Calculations. The potential energy surface for the gas-phase reaction of X- (X ) F, Cl) with HCOOCH3 was initially explored at the MP2 level of theory using the standard 6-31+G* basis set. This level of calculation was used to locate the structures corresponding to local minima and transition states. All stationary points were characterized by harmonic frequency analysis, and the frequencies were used to determine zero-point vibrational energies (ZPE) without any scaling procedure. For the F- + HCOOCH3 system, higher-level singlepoint calculations at the MP2/6-311+G(2df,2p) and MP4/631+G* levels were performed in conjunction with an additivity approximation to obtain effective MP4/6-311+G(2df,2p) energies. Comparison of the most significant results obtained by this methodology with those derived from calculations carried out with full MP4/6-311+G(2df,2p) energies reveal very good agreement and are discussed later in this paper. For the Cl- + HCOOCH3 system we have only used the MP2/6-31+G* energies since the relatively high barriers encountered for this system precluded an investigation at a higher level of theory for reasons explained in the discussion. All ab initio calculations were done with the Gaussian 94 suite of programs.23 The F- + HCOOCH3 System. Figure 2 outlines the detailed mechanisms of the reactions that were characterized in this work for the F- + HCOOCH3 system. Each step of the mechanism as well as intermediates and transition states has been numerically labeled for convenience. The optimized structures for all minima and transition states encountered for the different pathways are shown in Figures 3 and 4, while Table 1 lists the reaction and activation energies for each step. A graphical representation of the potential energy surface profile is shown in Figure 5. Minima and Transition State Structures. The interaction of F- with methyl formate leads to the formation of two ion-

Gas-Phase Reaction of Fl- and Cl- with Methyl Formate

J. Phys. Chem. A, Vol. 106, No. 2, 2002 373

Figure 2. A general diagram of the reaction pathways for the gas-phase F- + HCOOCH3 system.

Figure 3. Structure of intermediates and transition states resulting from the initial formation of the SN2 entrance complex (MS1) leading to the SN2 reaction and to formation of the tetrahedral species (MS3).

molecule complexes labeled MS1 and MS4 in Figures 2-5. The MS1 complex can undergo an SN2-type displacement through transition state TS2 that leads to the MS2 exit complex of the well-established SN2 double well potential.2 This complex can then dissociate to the final products, i.e., HCOO- and methyl fluoride. While passage of the MS1 to the MS2 complex is calculated to involve a barrier of 9.0 kcal mol-1, this transition state (TS2) is still well below the energy of the reagents. The calculated exothermicity for the SN2 reaction amounts to 23.7 kcal mol-1 in close agreement with a value of 21.5 ( 3 kcal mol-1derived from available experimental data.24 The MS1 complex can also proceed to form a tetrahedral intermediate (MS3 structure) through transition state TS4, and the barrier for this step is calculated to be only 4.1 kcal mol-1. This pathway is similar to that found for the gas-phase reaction of OH-.15,25 While the MS3 tetrahedral intermediate can undergo methanol elimination through step 8 to yield the MS5 complex, this step involves a barrier of 21.4 kcal mol-1 with the TS8 structure calculated to be 2.1 kcal mol-1 aboVe the energy of the reactants (see Figure 5).

Figure 4. Structure of intermediates and transition states resulting from abstraction of the formyl hydrogen by the fluoride ion.

By comparison, the MS4 ion-neutral complex can proceed either through steps 6 or 7 (Figures 2 and 5). Step 7 leads to the formation of the tetrahedral intermediate through transition state TS7 located 6.9 kcal mol-1 above the energy of the complex but 13.9 kcal mol-1 below the energy of the reagents. The fate of this tetrahedral intermediate was discussed above and further reaction is expected to be unfavorable on the basis of our calculations. The other pathways, step 6, leads to the proton abstraction of the formyl proton and formation of the three-body MS5 complex. Unlike the energy surface calculated for the similar OH- reaction,15 step 6 has a substantial energy barrier amounting to 11.2 kcal mol-1 at our best level of calculation.26 The three-body MS5 complex is characterized by a very low binding energy (3.6 kcal mol-1) with respect to the ultimate products of reaction 2a, namely the fluoride-methanol adduct, MS6, and CO. The binding energy of the fluoridemethanol is calculated to be 30.1 kcal mol-1, in excellent agreement with the experimental12a binding enthalpy of 29.6 (

374 J. Phys. Chem. A, Vol. 106, No. 2, 2002

Pliego and Riveros

TABLE 1: Reaction and Activation Energies for the Reaction Pathways of F- + HCOOCH3a Reaction Energies step

MP2/6-31+G(d)

MP2/6-311+G(2df,2p)

MP4/6-31+G(d)

MP4/6-311+G(2df,2p)b

∆ZPEc

∆E

1 2 3 4 5 6 7 8 9 10

-15.81 -13.67 11.50 0.68 -17.75 1.42 2.63 -1.21 3.72 28.56

-17.73 -15.45 11.07 -1.69 -20.14 0.77 0.72 0.05 3.59 30.77

-16.11 -13.59 11.82 0.57 -18.22 -0.19 2.68 -2.87 4.08 27.97

-18.04 -15.37 11.39 -1.80 -20.61 -0.84 0.77 -1.61 3.94 30.18

0.23 -1.13 -0.73 0.34 -0.19 -3.25 0.76 -4.01 -0.34 -0.13

-17.81 -16.50 10.66 -1.46 -20.80 -4.09 1.53 -5.62 3.61 30.05

step

MP2/6-31+G(d)

MP2/6-311+G(2df,2p)

MP4/6-31+G(d)

MP4/6-311+G(2df,2p)b

∆ZPE‡

∆E‡

2 4 6 7 8

11.75 4.54 20.04 6.76 27.77

12.00 4.54 15.41 6.95 25.05

9.81 4.14 19.04 6.36 27.14

10.07 4.15 14.42 6.55 24.42

-1.08 -0.06 -3.09 0.36 -3.06

8.99 4.09 11.33 6.91 21.36

Activation Energies

a Energies in units of kcal mol-1; steps numbered according to Figure 2. b Obtained by additivity approximation. c Zero-point vibrational energy contribution.

Figure 5. Calculated potential energy profile for the gas-phase F- + HCOOCH3 reaction.

2 kcal mol-1 and a value of 30.4 kcal mol-1 calculated previously at the CCSD level with a large basis set.27 Analysis of the Reaction Paths. The potential energy surface profile displayed in Figure 5 and discussed above is qualitatively consistent with the experimental results and with the nature of the reaction products in reaction 2. According to this profile, collisions between F- ions and methyl formate lead to the formation of MS1 and MS4 ion-molecule complexes. Both of these species can form the tetrahedral intermediate MS3 through TS4 or TS7, representing barriers of relatively low activation barriers when compared with the energy content of the system. However, this tetrahedral intermediate is unlikely to lead to reaction products since TS8 involves a high barrier as discussed above. However, this tetrahedral intermediate provides a communication channel between the MS1 and MS4 complex. Since ion-molecule reactions typically proceed through long-lived intermediates, it is possible to assume an equilibrium distribution

between the MS1, MS3, and MS4 complexes (with the MS4 complex being the predominant one) and the final products to be formed through TS2 and TS6. The efficiency of the SN2 reaction is thus expected to depend on the probability of passing through TS2 since the MS2 complex would be formed with enough energy for dissociation to the HCOO- + CH3F products. By comparison, the efficiency of the decarbonylation reaction will be dictated by the probability of passing through TS6, a pathway with a lower activation barrier than TS2. The final products are then expected to be CH3OH‚‚‚F- + CO as a result of dissociation of the energy-rich MS5 complexes. RRKM Calculations. The expected product distribution for reaction 2 can be estimated from a microcanonical analysis of the unimolecular rate constants for the intermediate complex. This approach assumes that the system displays statistical behavior, a reasonable starting point in the absence of any

Gas-Phase Reaction of Fl- and Cl- with Methyl Formate TABLE 2: Harmonic Vibrational Frequencies for the Structures MS4, TS2, and TS6a

a

MS4

TS2

TS6

108 109 170 232 331 415 749 933 1119 1188 1224 1242 1491 1505 1530 1551 1687 2582 3097 3179 3211

60 124 177 333 357 370 770 1049 1128 1162 1243 1348 1426 1437 1440 1700 2948 3238 3440 3444 -

53 77 170 195 262 294 532 728 799 1037 1175 1180 1469 1524 1547 1688 3033 3094 3135 3505 -

Units of cm-1. Values obtained at MP2/6-31+G(d) level of theory.

experimental data to the contrary. We have previously used this kind of approach in the analysis of the product distribution of the OH- + HCOOCH3 reaction and the theoretical predictions were found to be in excellent agreement with experimental observations.15 Thus, a simple RRKM approach was also used in the present case to estimate the product distribution for the F-/HCOOCH3 system. The analysis is based on the assumption that equilibrium between the MS1, MS3, and MS4 species is faster than the lifetime of these intermediate species. Under these conditions, we can take the MS4 complex, the most stable, as the reference species and compute the reaction rate constants through TS2 and TS6 by RRKM theory. For the sake of simplicity rotational effects have not been considered (J ) 0), and the internal energy available (E*) to the MS4 complex is then calculated by the following equation:

E* ) ∆EMS4 + ER - EMS4 where ∆EMS4 is the energy released upon formation of the MS4 complex (20.80 kcal mol-1), ER is the thermal contribution to the translational, rotational, and vibrational energy of the reactants (3.70 kcal mol-1), and EMS4 is the thermal contribution to the translational and rotational energy of the MS4 complex (1.78 kcal mol-1). Thus, at 298 K the energy available to the MS4 species is E* ) 22.7 kcal mol-1. The barrier heights for TS2 and TS6 are 12.0 kcal mol-1 and 11.3 kcal mol-1, respectively, and the calculated harmonic vibrational frequencies are listed in Table 2. RRKM calculations were then performed using the Zhu and Hase program,28 and the resulting unimolecular rate constants (kTS2 ) 4.70 × 1010 s-1 and kTS6 ) 5.08 × 1011 s-1) yield the following branching ratio for reaction 2: 92% decarbonylation (the Riveros reaction 2a) and 8% for the SN2 path (reaction 2b). This result is in very good agreement with our previous measurements7 and with those reported in the present set of experiments. Furthermore, this branching ratio remains essentially unchanged with increasing internal energy of the reagents as verified by RRKM calculations carried out with higher values of E*.29 This prediction also agrees with the experiments that show no significant variation with the thermalization time of the ions.

J. Phys. Chem. A, Vol. 106, No. 2, 2002 375 TABLE 3: Single-Point Calculations at Exact MP4/ 6-311+G(2df,2p) and MP2/6-311+G(2df,2p) Levels and Using Additivity Approximation for the MP4 Methoda process

MP4 (exact)

MP4 (additivity)b

MP2

MS4 f TS2 MS4 f TS6 TS2 f TS6

12.38 14.03 1.65

12.64 14.42 1.78

14.41 15.41 1.00

a In kcal mol-1. b In the additivity approximation E[MP4/6311+G(2df,2p)] ≈ E[MP4/6-31+G(d)] + E[MP2/6-311+G(2df,2p)] - E[MP2/6-31+G(d)].

A similar RRKM calculation was also carried out for the F+ DCOOCH3 reaction by using the calculated vibrational frequencies of the deuterium-substituted MS4 and the corresponding transition state structures. The resulting unimolecular rate constants, kTS2(DCOOCH3) ) 4.0 × 1010 s-1 and kTS6(DCOOCH3) ) 3.9 × 1011 s-1, predict for DCOOCH3 a branching ratio of 91% for the decarbonylation path (the Riveros reaction 2a) and 9% for the SN2 path (reaction 2b). This small change in branching ratio upon deuteriation at the formyl position is also in excellent agreement with the present experimental findings. Due to the importance of the local energy barriers in the overall energy surface and the barrier heights in determining the calculated branching ratio, we have tested the accuracy of our additivity approximation by performing exact MP4/6311+G(2df, 2p) single-point calculations on the key structures, namely MS4, TS2, and TS6. Table 3 lists the barriers for the MS4 f TS2 and MS4 f TS6 processes, as well as the energy variation involved in going from TS2 to TS6, calculated at the MP4/6-311+G(2df,2p) level, at the MP2/6-311+G(2df,2p) level, and using the additivity approximation. The results shown in Table 3 clearly point out that the additivity approximation is indeed very adequate with barrier heights that are predicted to be different by less than 0.4 kcal mol-1 with those calculated at the exact MP4/6-311+G(2df,2p) level. By comparison, barrier heights calculated at the MP2/6-311+G(2df,2p) level differ as much as 2 kcal mol-1 from those calculated at higher levels of theory. Even more important and critical to the calculation of the branching ratio is the calculated variation in energy in going from TS2 to TS6 structure (TS2 f TS6). The additivity approximation yields a value that is only 0.13 kcal mol-1 different from that obtained at the full MP4 calculation, while the MP2 calculation yields a value that differs by 0.65 kcal mol-1. These results clearly point out that the additivity approximation used in our work is capable of predicting a potential energy surface in very close agreement with that obtained at the exact MP4 level and at a considerable savings of computational time. Thus, very small differences in the calculated branching ratio would result from using the values obtained by the exact MP4 calculation or by additivity approximation. On the other hand, the MP2 calculations would lead to a considerably less reliable potential energy surface and branching ratio. The Cl- + HCOOCH3 System. The fact that reaction 1a is not observed with Cl- led us to analyze the profile of the potential energy surface in a fashion similar to that described above. The available experimental data20 predicts the Cl-promoted SN2 reaction (reaction 1b in Scheme 1) to be endothermic by 9 ( 3 kcal mol-1. Thus, the fact that no HCOOdisplacement is observed in the ion-molecule reaction of Clwith HCOOCH3 can be readily understood on thermochemical grounds. On the other hand, similar experimental data30 predict the so-called Riveros (reaction 1a) to be exothermic by 6 ( 2 kcal mol-1.

376 J. Phys. Chem. A, Vol. 106, No. 2, 2002

Pliego and Riveros

Figure 6. A general diagram of the reaction pathways for the gas-phase Cl- + HCOOCH3 system.

TABLE 4: Reaction and Activation Energies for the Gas-Phase Reaction Cl- + HCOOCH3a Reaction Energies step

MP2/6-31+G*

∆ZPEb

∆E

1 2 3 4 5 6 7 im

-9.67 6.91 11.97 -9.30 5.52 3.58 16.16 32.41

0.29 -2.13 -0.65 0.32 -2.89 -0.56 -0.78 0.00

-9.38 4.78 11.32 -8.98 2.63 3.03 15.38 32.41

Activation Energies step

MP2/6-31+G*

∆ZPE‡

∆E‡

2 5

22.69 49.53

-1.82 -5.56

20.87 43.97

a Energies in units of kcal mol-1; steps numbered according to Figure 6. b Zero-point vibrational energy contribution.

Figure 7. Structure of intermediates and transition states resulting from attachment of the chloride ion to the methyl group of the ester.

Our theoretical results for this system are shown in Figures 6 to 9 and are summarized in Table 4. Figure 6 outlines the reaction pathways investigated by analogy with the F- reactions, and Figures 7 and 8 display the optimized structures for the local minima and transition states. The calculated potential energy surface profile is represented in Figure 9. Analysis of the Potential Energy Surface. As in the case of the fluoride ion, the interaction of the Cl- ion with methyl formate leads to two possible ion-molecule complexes, namely MS1 and MS3. These are calculated to have stabilization energies close to 9 kcal mol-1. The SN2 mechanism for the reactant side MS1 complex is predicted to have a substantial

Figure 8. Structure of intermediates and transition states resulting from abstraction of the formyl hydrogen by the chloride ion. The MSi species is a putative tetrahedral intermediate that is not a minimum in the potential energy surface (see text).

activation energy and the TS2 structure is located 11.5 kcal mol-1 above the reagents. Thus, passage to the exit MS2 complex, located 4.6 kcal mol-1 below the energy of the reagents, and dissociation of MS2 to HCOO- and methyl chloride is highly unfavorable at thermal energies. The overall endothermicity of the SN2 reaction is predicted to be 6.7 kcal mol-1 at the MP2/6-31+G* level of theory. The decarbonylation mechanism involving the MS3 complex was found to have a very high activation energy associated with the TS5 structure, located 35 kcal mol-1 above the energy of the reagents! This barrier allows us to conclude that no reaction is expected to be observed by techniques such as ion-cyclotron resonance or flowing afterglow even though the overall reaction is exothermic. The potential energy surface for this pathway also reveals the three-body product side complex, MS4 in Figure 9, that is located 6.4 kcal mol-1 below the energy of the reagents. If this species were to be formed, it would then lead to loss of CO and formation of the CH3OH‚‚‚Cl- adduct. Another possibility that we have investigated is formation of a tetrahedral intermediate as encountered for the F- ion. In this case, the tetrahedral species is found not to be minimum on the potential energy surface. To estimate the stability of a hypothetical tetrahedral like structure for this system, a calculation was performed with the C-Cl distance frozen at 1.8 Å, and optimizing the remaining geometrical parameters. This is

Gas-Phase Reaction of Fl- and Cl- with Methyl Formate

J. Phys. Chem. A, Vol. 106, No. 2, 2002 377

Figure 9. Calculated potential energy profile for the gas-phase Cl- + HCOOCH3 reaction.

represented by the MSi structure located 23.0 kcal mol-1 above the energy of the initial reagents. This energy surface profile clearly predicts that both channels are energetically unfavorable and are consistent with the experimental results that show no reactivity for this system.16 The fact that the barrier for the decarbonylation reaction was found to be very high precluded the need for calculations at higher levels of theory. On the basis of the more extended calculations for the F-/HCOOCH3 system, we can in fact estimate that the barrier predicted by our MP2/6-31+G(d) calculation for the Cl-/HCOOCH3 reaction could probably be in error by 2 or 3 kcal mol-1 at most. Even with this uncertainty, the transition structure TS5 would still lie considerably above the energy of the reagents, and no reaction is expected to be observed within the conventional time and pressure regimes of gas-phase ion chemistry techniques. Discussion The potential energy surface for the gas-phase reaction of anionic nucleophiles with methyl formate, discussed in this paper and in ref 15, reveals the formation of a highly favorable complex with the anion bound to the formyl hydrogen of the ester. Our calculations show that the barrier associated with further abstraction of this proton and the subsequent formation of an anion-methanol adduct and CO scales with the basicity of the corresponding nucleophile. In fact, these studies clearly show the barrier to increase in the order OH- (no barrier)15 < F- , Cl-, with the last case involving an energy barrier well above that of the reagents. A second reaction complex that can also be formed in these systems is that corresponding to the SN2 entrance complex.31 The activation energy calculated for these complexes to proceed to HCOO- and MeX (X- ) nucleophile) also follows the same trend with OH- < F- , Cl-, although in the latter case the overall reaction is endothermic. A second important aspect is the question of nucleophilic attack at the carbonyl center and formation of a tetrahedral species in a simple ester such as methyl formate. For OH- and F-, the potential energy surface clearly shows that formation of a stable tetrahedral intermediate in these cases stems from

the complexes discussed above, and not as a direct process. For the Cl- ion, the tetrahedral species of methyl formate is not a stable intermediate but a transition state, a result that is of considerable significance. An earlier theoretical study of the thermoneutral chloride displacement reaction between Cl- and either formyl chloride, HCOCl, or acetyl chloride, CH3COCl, predicted the tetrahedral structures to be transition states for these reactions rather than stable intermediates.32 These findings are consistent with experimental observations that the chlorines are not equivalent in the ion-neutral complexes of Cl- with CH3COCl and CH3OCOCl, and thus these complexes are best described as ion-dipole complexes rather than tetrahedral intermediates.33 Finally, we have made use of a very simple approach to estimate the product distribution in these systems. This approach, based on RRKM theory, predicts branching ratios that are in excellent agreement with experimental findings. The significant changes in the potential energy surface for equivalent reactions in solution is presently being investigated in order to learn how solvent effects can extensively modify intrinsic reactivity and the outcome of well-known chemical reactions. Acknowledgment. This work was made possible by the support of a grant and a postdoctoral fellowship (J.R.P.) from the Sa˜o Paulo Science Foundation (FAPESP), and a Senior Research fellowship from the Brazilian Research Council (C.N.Pq). The authors also acknowledge the help of Gustavo H. Guedes (undergraduate FAPESP scholarship) with some of the experiments. References and Notes (1) For two recent reviews, see (a) Gronert, S. Chem. ReV. 2001, 101, 329. (b) DePuy, C. H. Int. J. Mass Spectrom. 2000, 200, 79. (2) For a general discussion of potential energy surfaces for these reactions, see Chabinyc, M. L.; Craig, S. L.; Regan, C. K.; Brauman, J. I. Science 1998, 279, 1882, and references therein. (3) For a benchmark large-scale coupled cluster calculation for the F+ CH3Cl reaction, see Botschwina, P.; Horn, M.; Seeger, S.; Oswald, R. Ber. Bunsen-Ges. Phys. Chem. 1997, 101, 387. (4) Hase, W. L. Science 1994, 266, 998.

378 J. Phys. Chem. A, Vol. 106, No. 2, 2002 (5) (a) Jose, S. M.; Riveros, J. M. NouVeau J. Chim. 1977, 1, 113. (b) Takashima, K.; Riveros, J. M. J. Am. Chem. Soc. 1978, 100, 6128. (c) Wilbur, J. L.; Brauman, J. I. J. Am. Chem. Soc. 1994, 116, 12125. (d) DePuy, C. H.; Grabowski, J. J.; Bierbaum, V. M.; Ingemann, S.; Nibbering, N. M. M. J. Am. Chem. Soc. 1985, 107, 1093. (e) Bernasconi, C. F.; Stronach, M. W.; DePuy, C. H.; Gronert, S. J. Am. Chem. Soc. 1990, 112, 9044. (f) Frink, B. T.; Hadad, C. M. J. Chem. Soc., Perkin Trans. 2 1999, 2397. (g) van der Wel, H.; Nibbering, N. M. M. Recl. TraV. Chim. Pays-Bas 1988, 107, 479. (6) Blair, L. K.; Isolani, P. C.; Riveros, J. M. J. Am. Chem. Soc. 1973, 95, 1057. (7) Faigle, J. F. G.; Isolani, P. C.; Riveros, J. M. J. Am. Chem. Soc. 1976, 98, 2049. (8) van der Wel, H.; Nibbering, N. M. M. Int. J. Mass Spectrom. Ion Processes 1986, 72, 145. (9) Johlman, C. L.; Wilkins, C. L. J. Am. Chem. Soc. 1985, 107, 327. (10) Reaction 1a has been informally referred to in gas-phase ion chemistry as the “Riveros reaction”. (11) Takashima, K.; Riveros, J. M. Mass Spectrom. ReV. 1998, 17, 409. (12) (a) Larson, J. W.; McMahon, T. B. J. Am. Chem. Soc. 1983, 105, 2944. (b) Caldwell, G.; Rozeboom, M. D.; Kiplinger, J. P.; Bartmess, J. E. J. Am. Chem. Soc. 1984, 106, 4660-4667. (c) Chabinyc, M. L.; Brauman, J. I. J. Am. Chem. Soc. 1998, 120, 10863. (13) (a) Chabinyc, M. L.; Brauman, J. I. J. Am. Chem. Soc. 2000, 122, 8739. (b) Baer, S.; Brauman, J. I. J. Am. Chem. Soc. 1992, 114, 5733. (c) Dodd, J. A.; Baer, S.; Moylan, C. R.; Brauman, J. I. J. Am. Chem. Soc. 1991, 113, 5942. (d) Lim, K. F.; Brauman, J. I. J. Chem. Phys. 1991, 95, 7164. (14) (a) Yang, Y.; Linnert, H. V.; Riveros, J. M.; Williams, K. R.; Eyler, J. R. J. Phys. Chem. A 1997, 101, 2371. (b) Peiris, D. M.; Riveros, J. M.; Eyler, J. R. Int. J. Mass Spectrom. Ion Processes 1996, 159, 169. (c) Mihalick, J. E.; Gatev, G. G.; Brauman, J. I. J. Am. Chem. Soc. 1996, 118, 12424. (15) Pliego, J. R., Jr.; Riveros, J. M. Chem. Eur. J. 2001, 7, 169. (16) Isolani, P. C.; Riveros, J. M. Chem. Phys. Lett. 1975, 33, 361. (17) (a) Silva, M. L. P.; Riveros, J. M. J. Mass Spectrom. 1995, 30, 733. (b) Morgon, N. H.; Argenton, A. B.; Silva, M. L. P.; Riveros, J. M. J. Am. Chem. Soc. 1997, 119, 1708. (c) Morgon, N. H.; Xavier, L. A.; Riveros, J. M. Int. J. Mass Spectrom. 2000, 195, 363. (d) Moraes, P. R. P.; Linnert, H. V.; Aschi, M.; Riveros, J. M. J. Am. Chem. Soc. 2000, 122, 10133. (18) This methodology has been shown to yield reliable kinetic data. For some recent examples, see (a) Nixdorf, A.; Gru¨tzmacher, H. F. Chem. Eur. J. 2001, 7, 1248. (b) Bu¨chner, M.; Gru¨tzmacher, H. F. Int. J. Mass Spectrom. 2000, 199, 141. (c) Nixdorf, A.; Gru¨tzmacher, H. F. Int. J. Mass Spectrom. 2000, 195/196, 533.

Pliego and Riveros (19) McIver, R. T.; Hunter, R. L.; Baykutt, G. Anal. Chem. 1989, 61, 491. (20) For some of the merits of impulse excitation, see Uechi, G. T.; Dunbar, R. C. J. Am. Soc. Mass Spectrom. 1992, 3, 734. (21) For problems associated with quantitation of ionic products under chirp excitation, see, (a) Rempel, D. L.; Huang, S. K.; Gross, M. Int. J. Mass Spectrom. Ion Processes 1986, 70, 163. (b) Moini, M.; Eyler, J. R. Int. J. Mass Spectrom. Ion Processes 1989, 87, 29. (22) Baer, S.; Stoutland, P. O.; Brauman, J. I. J. Am. Chem. Soc. 1989, 111, 4097. (23) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.; Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, G. A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski, V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; HeadGordon, M.; Gonzalez, C.; Pople, J. A. Gaussian 94, Version D.2, Revision D.2; Gaussian, Inc.: Pittsburgh, PA, 1994. (24) NIST Chemistry WebBook, NIST Standard Reference Database Number 69; Mallard, W. G., Lindstrom, P. J., Eds.; NIST: Gaithersburg, MD, 2000; (http://webbook.nist.gov/chemistry/). (25) Zhan, C.-G.; Landry, D. W.; Ornstein, R. L. J. Am. Chem. Soc. 2000, 122, 1522. (26) The barrier of this step is sensitive to the size of the basis set and the level of correlation energy included in the calculation. In fact, as reported in ref 5f, calculations at the B3LYP/6-31+G* level predict this step to involve a barrier that is slightly above the energy of the reagent molecules! (27) Wladkowski, B. D.; East, A. L. L.; Mihalick, J. E.; Allen, W. D.; Brauman, J. I. J. Chem. Phys. 1994, 100, 2058. (28) Zhu, L.; Hase, W. L. A General RRKM Program; Department of Chemistry, Wayne State University, 1993. (29) Calculations were performed up to 10 kcal mol-1 of excess energy. (30) This calculation is based on the revised binding energy for the chloride ion-methanol adduct as reported by Bogdanov, B.; Peschke, M.; Tonner, D. S.; Szulejko, J. E.; McMahon, T. B. Int. J. Mass Spectrom. 1999, 185/186/187, 707. (31) (a) Riveros, J. M.; Breda, A. C.; Blair, L. K. J. Am. Chem. Soc. 1973, 95, 4066. (b) Riveros, J. M.; Jose, S. M.; Takashima, K. AdV. Phys. Org. Chem. 1985, 21, 197. (32) Blake, J. F.; Jorgensen, W. L. J. Am. Chem. Soc. 1987, 109, 3856. (33) (a) Asubiojo, O. I.; Brauman, J. I. J. Am. Chem. Soc. 1979, 101, 3715. (b) Wilbur, J. L.; Brauman, J. I. J. Am. Chem. Soc. 1994, 116, 5839.